June 22, 2011
HAMBURG, Germany, June 22 -- CAPS entreprise and Pathscale today announced that OpenHMPP - a non-profit association of academic, research, and industry partners which mission is to work with and for the HPC community to make HMPP directives a world-class open standard -- is now live!
Since last year, CAPS and Pathscale have been coordinating their effort to make HMPP directives an Open Standard denoted OpenHMPP. This initiative aims at creating a community effort for proposing a directive manycore programming interface and thus capitalizing on existing experiences with HMPP.
"CAPS has always been committed in delivering to users solutions that will secure their software investment. We think the directive approach is definitely the best way to meet this issue. Now, with the open standard and the software companies adopting HMPP, our customers feel confident in their choice." declares Laurent Bertaux, CAPS CEO. The HMPP directives, initially developed by CAPS, have been designed and used for more than 3 years by major HPC leaders. As a high level of abstraction for programming GPUs in scientific applications, HMPP preserves customers' software assets by keeping applications portable. The HMPP directive-based programming model offers a powerful syntax to efficiently offload computations on hardware accelerators and to optimize data movement.
HMPP directives describe remote procedure call (RPC) on an accelerator device (e.g. GPU) or more generally a set of cores. The directives annotate C or Fortran codes to describe two sets of functionalities: the offloading of procedures (denoted codelets) onto a remote device, the optimization of data transfers between the CPU main memory and the accelerator memory.
As more and more industry leaders, ALLINEA Debugger now supports HMPP directives. "With the recent release of Allinea DDT 3.0, it is now possible to debug your NVIDIA GPU application at a high level", says Jacques Philouze, ALLINEA Vice President Sales & Marketing.
The OpenHMPP consortium oversees the HMPP directives specification, produces and approves new versions of the specification. In accordance with its vision involving participation and knowledge sharing, OpenHMPP works towards providing programmers with a portable, scalable directive-based programming model for developing parallel applications on manycore platforms.
OpenHMPP group has several top-level goals:
A set of core partners (CAPS, GENCI, ICHEC, INRIA, PathScale…) are currently in discussion to set up the non-profit organization that will own the OpenHMPP brand and run its evolutions. This organization will gather technology providers, application developers as well as ISVs.
"It's exciting for PathScale to be a founding member in the OpenHMPP organization." Comments Christopher Berström, CTO at Pathscale. "Almost two years ago Francois Bodin (CTO at CAPS) and I started the idea to make the HMPP directives an open standard, but now a clear path for broad collaboration and adoption is possible for anyone interested."
About CAPS entreprise
CAPS is a major supplier of solutions dedicated to application deployment on manycore processors with a focus on GPUs. CAPS develops HMPP (Heterogeneous Multicore Parallel Programming), a hybrid compiler that gives developers an easy access to manycore systems and offers services ranging from manycore parallel programming trainings to complete application porting and expertise. Website: www.caps-entreprise.com
PathScale Inc. has developed industry leading high performance Fortran, C and C++ compiler products for AMD64, Intel 64, MIPS processors and provides support to users desiring the highest level of performance from their applications. The PathScale EKOPath Compiler Suite has the world's most advanced optimization infrastructure and can fully exploit the potentials of many-core architectures. The company's goal is to deliver robust and high performance compilers tailored to clustered, GPGPU and multi-core computing environments. More information about PathScale is available on the web at http://www.pathscale.com.
Source: CAPS entreprise; PathScale
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?